AI Sophistication and Big Data Organisational Culture as Dual Drivers of Customer Experience and Public Trust in UAE Government Services

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Scientific Problem: Governments worldwide are deploying AI-enabled services at scale, yet uncertainty persists about whether these systems genuinely contribute to citizen satisfaction and institutional trust and what organisational conditions make them do so. Existing research has examined technical AI capability and organisational data culture in isolation, leaving unresolved whether both exert independent effects when modelled together and whether emotional and relational mechanisms explain how these antecedents reach citizen outcomes. Aim: This study tests whether AI Sophistication (AIS) and Big Data Organisational Culture (BDOC) independently influence Customer Experience (CX) and Public Trust (PT) in UAE government AI services, whether Emotional Intelligence (EI) and the Real-Time Commensality (RTC) scale developed in this study mediate those relationships, and whether Multicultural Competence (MC) moderates them. Method: Survey data were collected between October and December 2024 from N = 500 verified users of UAE government AI services across healthcare, education, transportation, and administrative domains. All seven constructs were validated against standard psychometric thresholds (AVE > 0.70, CR > 0.88, HTMT < 0.85). A Dual-Driver Mediation–Moderation Model was tested using hierarchical regression, bootstrapped mediation (Hayes PROCESS macro, 5000 resamples), and interaction-term moderation. Results: AIS and BDOC showed near-identical associations with CX (B = 0.567 and B = 0.568, p < 0.001), consistent with dual-driver equivalence. In the full model, BDOC substantially outperformed AIS as a predictor of PT (B = 0.294 vs B = 0.108). All eight mediation pathways were supported, with EI showing consistently larger indirect effects (range: 0.070–0.093) than RTC (range: 0.049–0.072). EI recorded the lowest construct mean (M = 1.959) despite being the strongest mediator—the affective gap. MC did not moderate any pathway, indicating no model-level evidence of differential associations across levels of Multicultural Competence; given the configural-only invariance assessment, this should not be interpreted as conclusive proof of cultural convergence. Conclusions: These findings extend the DeLone and McLean IS Success Model by positioning BDOC as a co-equal predictor of citizen satisfaction and the stronger predictor of institutional trust—suggesting that sustainable government AI delivery may depend as much on organisational culture as on technical capability. Future Research: Longitudinal designs, cross-country comparisons, and validation of RTC against established social presence measures are identified as priority future directions.